The present proposal is concerned with the prediction of individual placebo responses in clinical samples. Here we part from the usual view of placebos as inert control states. On the contrary, it is hypothesized that placebo administration triggers a cascade of events activating endogenous mechanisms that promote homeostasis. This leads to the proposal that a substantial proportion of the variance in placebo responding in CNS trials can be explained by the functional variation of specific neurobiological circuits and mechanisms. We focus on the prediction of placebo effects in patients diagnosed with Major Depressive Episode (MDE), unmedicated at the time of the study. This illness was selected because of its high frequency and chronicity in the general population, but also one that presents with high placebo responses in controlled trials. Half of the subjects will also present with a diagnosis of nicotine dependence, which is expected to reduce placebo responses rates, but also affect the underlying neurobiology, increasing the generalizability of the findings. Non-problem alcohol use will be permitted and entered as a covariate in the analyses. A 3-step process is proposed. Studies using positron emission tomograph will determine the placebo-induced activation of neurotransmitter systems through to be involved in both the pathophysiology of MDE and the effects of expectations. Functional magnetic resonance imaging will be employed to determine the proportion of the variance in placebo effects explained by the function of reward, decision-making and "motivation" regions. Individual variations in neurotransmitter systems and circuits involved will then be modeled by a combination of neuropsychological tests and the presence of common genetic polymorphisms regulating those regional networks. This proposal therefore addresses the predictability of the placebo effect and its underlying neurobiology. The capacity to utilize internal resources to change clinical conditions (as opposed to traditional therapies that are given or applied to the patient with little individual control) represents both a shift in paradigm and a source of "noise" in clinical trials. As such, the results of the studies proposed have the potential for lasting impact in the practice of medicine at large.